Hacker News Reader

Comment by smogcutter | Hacker News Reader

That okcupid stats blog was great, here’s a thread on reddit where someone archived them: https://www.reddit.com/r/gwern/comments/aapn1l/okcupid_blog_...

To sum up for parent: in okcupid’s data, women considered attractive but not super hot got fewer messages than women who were less conventionally attractive but had more polarizing photos. In other words for most women it’s a better strategy to play up a look/feature that’s unique but potentially alienating than it is to be generically attractive.

View on HN

Polarizing being good for dating doesn't necessarily mean it's good for influencers. In dating you want one person to like you a lot. So if you make the majority dislike a bit and the minority like you more that's great. But in influencing you want a lot of people to like you a little.

That’s one explanation. Another possibility is a more polarizing face makes some guys feel like they have a better chance. This is total speculation, of course, just offering up another way to explain the data. Could be a mix of both explanations.

I went back and read the post, and I remembered it wrong. The effect is there for women with the same overall attractiveness rating, which is a stronger version than I originally posted.

> Another possibility is a more polarizing face makes some guys feel like they have a better chance.

This explains why guys would send messages, but not why they would rate pics highly. It’s a better fit for why guys would message girls with lower ratings, not girls with high ratings and high std. dev. The trick in their version is it’s not just “I think other guys aren’t into her”, it’s “I think other guys aren’t into her, but I am.”

But yeah, being a story that explains the data doesn’t make it true, and all sorts of stuff could be going on. And like the other poster pointed out their analysis is fun but not necessarily ironclad. For my part I figure there’s a ton of cultural signaling going on in dating pics that they can’t really account for.

I loved reading OkCupid's Blog but it rarely matched my experience with the site and the conclusions they reached almost always had other plausible explanations.

Reply on news.ycombinator.com